firestorm
firestorm copied to clipboard
deal with spam
https://forum.firestormforum.org/categories/8/threads/11#post-49
it was bound to happen sooner or later. I was hoping later. we should gather prior art / how other systems do this. I'm a fan of reverse captcha generally, maybe we just start with that?
I wish the Google reCAPTCHA was automatic. I was just in gmail, clicked a link to a forum, Google domain (gmail), plus another Google domain(reCAPTCHA), link this stuff and clear me straight away....
I've dealt with a small amount of spammy concerns, have a few suggestions:
-
Use Google Invisible Captcha (requires privacy policy update) on content-creation pages, including new post, new comment, new user account creation, etc etc. Log on your side whenever Google tells you an action failed the human-check so you can build a forum-specific "suspicious actions" dataset, such as spammy IP addresses, words or content, etc.
-
Log account creation stats - e.g. account_creation_ip, account_creation_country, for later use in identifying manually-created accounts that will later be used either by bots or humans. For example, users typically create either a post or a comment very soon after creating an account; bots either immediately spam or lie dormant.
-
Allow users to flag comments as spam, but as with any user-input, trust but verify.
I think it used to be in some Forums, even recent ones, that you'd have a human mod doing the greylist filter. New user A gets to verify their login data. They get sorted into a level of rights where they don't get to post, just read(?). After a certain amount of time spent reading, they get whitelisted (by a Mod) and now they can post replies. (optionally on top of that: after some amount of replies, they can open new topics)